The main objective of this paper is to discuss various deep learning models. Deep learning algorithms are used in multiple data science fields but play a vital role in image processing, diagnosing, and predicting. In recent days, the size of data, especially in terms of images, has increased rapidly with dimensions. Earlier research works used Machine learning algorithms, but they got down while analyzing a large volume of data. Most models have been created based on deep learning methods recently, incorporating data-driven models into data learning models by eliminating human interactions. Deep learning models learn the data in-depth to identify, detect, and classify medical image data. According to domain-specific learning, extracting hierarchical and deep learning models can do high-dimensional data. Thus, deep learning models prove themselves rapidly as state-of-the-art models for various emerging applications. One of the applications is medical image diagnosing applications. This paper mainly discussed medical image diagnosis using various deep-learning algorithms. It also highlights the success stories in data/image analytics, providing reassurance about the effectiveness of deep learning models. From the experimental results, it is verified that CNN (Convolution Neural Network) obtained an average accuracy value of 98.99%, which is higher than that of other algorithms. Also, CNN consumes a processing time of 863ms, less than the other algorithms. This paper concludes by discussing research issues and future guidelines for later improvement, highlighting the potential of this field for further research and development.